There is considerable interest in the field in interrogating genetic variants with pleiotropic effects [Giambartolomei et al., 2014; Hu et al., 2013; Lee et al., 2013; Tang and Lin, 2014, 2013], examining if genetic effects vary across cohorts/ethnic groups [Wen and Stephens, 2014], performing meta‐analyses that combine results from multiple studies [Liu et al., 2014] or implementing Mendelian randomization experiments by joint analyses of genetic associations with risk factors and disease outcomes [Do et al., 2013; Voight et al., 2012]. These research questions all require joint analysis of multiple sets of summary association statistics and their covariance information. Multiple studies may not have the same set of genetic variants genotyped (particularly for sequencing studies, where different variant sites are called in each study). It can be a nontrivial task to randomly access a large number of files of summary association statistics and covariance matrices, efficiently retrieve information specific to a genetic region of interest, and collate variant sites between studies. A great amount of ad hoc scripting may be needed.
